F or good reasons, economists have had a long-standing preference for studying peoples' revealed preferences; that is, looking at individuals' actual choices and decisions rather than their stated intentions or subjective reports of likes and dislikes. Yet people often make choices that bear a mixed relationship to their own happiness. A large literature from behavioral economics and psychology finds that people often make inconsistent choices, fail to learn from experience, exhibit reluctance to trade, base their own satisfaction on how their situation compares with the satisfaction of others and depart from the standard model of the rational economic agent in other ways. If people display bounded rationality when it comes to maximizing utility, then their choices do not necessarily reflect their "true" preferences, and an exclusive reliance on choices to infer what people desire loses some of its appeal. Direct reports of subjective well-being may have a useful role in the measurement of consumer preferences and social welfare, if they can be done in a credible way. Indeed, economists have already made much use of subjective well-being data. From 2001 to 2005, more than 100 papers were written analyzing data on selfreported life satisfaction or happiness, according to a tabulation of EconLit, up from just four in 1991-1995. Data on subjective well-being have been used by economists to examine both macro-and micro-oriented questions. In a classic paper, Journal of Economic Yet another use of subjective well-being has been to provide an external check on economic indicators. For example, Nordhaus (1998) and What are economists to make of this enterprise? Can well-being be measured by a subjective survey, even approximately? In this paper, we discuss research on how individuals' responses to subjective well-being questions vary with their circumstances and other factors. We will argue that it is fruitful to distinguish among different conceptions of utility rather than presume to measure a single, unifying concept that motivates all human choices and registers all relevant feelings and experiences. While various measures of well-being are useful for some purposes, it is important to recognize that subjective well-being measures features of individuals' perceptions of their experiences, not their utility as economists typically conceive of it. Those perceptions are a more accurate gauge of actual feelings if they are reported closer to the time of, and in direct reference to, the actual experience. We conclude by proposing the U-index, a misery index of sorts, which measures the proportion of time that people spend in an unpleasant state, and has the virtue of not requiring a cardinal conception of individuals' feelings. Measuring Subjective Experience in Principle and in the Lab The earliest popular conceptions of utility, from Jeremy Bentham through Francis Ysidro Edgeworth and Alfred Marshall, was as a continuous hedonic flow of pleasure or pain. Kahneman has called this conception experienced utility, and it is also similar to what Juster, Courant and Dow (1985) call process benefits. 1 Edgeworth 1 Juster, Courant and Dow define process benefits as the "direct subjective consequences from engaging in some activities to the exclusion of others. . . . For instance, how much an individual likes or dislikes the activity 'painting one's house,' in conjunction with the amount of time one spends in painting the house, is as important determinant of well-being independent of how satisfied one feels about having a freshly painted house." Journal of Economic Perspectives defined the happiness of an individual during a period of time as the sum of the momentary utilities over that time period; that is, the temporal integral of momentary utility. Several methods have been used to attempt to measure the moment-tomoment flow of pleasure or pain in the laboratory. An advantage of laboratory experiments is that extraneous aspects of an experience can be controlled, and the unique effect of a stimulus on individuals' experiences can be evaluated. Participants in many experiments in psychology and in consumer research, for example, are required to undergo an experience, such as being exposed to loud noises or watching a film clip. They are asked to provide a continuous indication of the hedonic quality of their experience in real time by manipulating a lever that controls a marker on a scale, which is usually defined by extreme values such as very pleasant and very unpleasant and by a neutral value. In a similar fashion, public opinion during a political debate is sometimes assessed by means of a "dial group," in which a group of observers continually indicate their pleasure or displeasure with the candidates' views by continuously adjusting a dial. These studies yield a temporal profile of moment-to-moment subjective experience. The results of these experiments provide some insights into more standard measures of satisfaction, and lend support for the usefulness of making a distinction between experienced utility and remembered utility; that is, the way people feel about experiences in real-time and the way they remember their experiences after they are over. The participants in experiments in which a physical stimulus is varied generally provide profiles that are similar, both in level and in shape, and that respond to the stimulus in a sensible way. For example, in an experiment described in Kahneman, Fredrickson, Schreiber and Redelmeier (1993), participant ratings of pain on a 0 -14 scale increased from an average of 4.0 (after the first ten seconds) to an average of 8.4 after 60 seconds of immersing their hand in water at 14°Celsius, and the average dropped to about 6.50 over the next 30 seconds, as the temperature of the water was gradually raised to 15°C. At the end of an experiment, individuals can be asked to evaluate their experience as a whole. Such a retrospective report can be thought of as representing the respondents' remembered utility. The evaluation of remembered utility requires the individual to remember a stream of experiences and to aggregate them in some way. Ideally, one would hope that the individual who reports his or her overall remembered utility for a period performs the task of summing momentary utilities over time that Edgeworth had in mind. This is not the case, however. Numerous studies have related individuals' retrospective evaluations of an experience to their record of real-time reports. Although retrospective evaluations are related to the real-time reports-people are generally correct in classifying a past episode as pleasant or awful-retrospective reports are also susceptible to systematic biases. Kahneman, Fredrickson, Schreiber and Redelmeier (1993) showed that retrospective evaluations of past episodes have the dimensionality of a weighted average of moment utilities, where the weights are not equal. The critical finding was that Daniel Kahneman and Alan B. Krueger 5 the duration of episodes of pain or discomfort was completely neglected in retrospective evaluations, and great weight was placed on the end of the experience and the peak or trough. Further evidence suggests that individuals' choices are affected by their remembered utility (as discussed by Kahneman and Thaler in this issue), not the profile of their experiences. This line of research suggests that the intensity of pain and pleasure that arises during an experience can be plausibly measured in real time and that retrospective assessments are not necessarily a good measure of the sum total of individuals' actual experiences. These observations are relevant because past research on well-being has relied almost entirely on reports of life satisfaction and happiness, which are global retrospective assessments. After evaluating research on global retrospective assessments of subjective well-being, we introduce alternative measures that attempt to tap experienced utility more directly. Surveys of Life Satisfaction and Happiness The questions most frequently asked in research using surveys of subjective well-being elicit reports of global life satisfaction or happiness. In the World Values Survey, for example, respondents in 81 countries are asked, "All things considered, how satisfied are you with your life as a whole these days?" The General Social Survey (GSS) similarly asks Americans, "Taken all together, how would you say things are these days? Would you say that you are very happy, pretty happy, or not too happy?" Respondents have little trouble answering these questions. In the 1998 GSS, for example, less than 1 percent of respondents refused to provide an answer or answered "don't know"; by contrast, 17 percent of respondents refused to provide their earnings. How should a social scientist interpret answers to questions about global life satisfaction or happiness? After all, life satisfaction is neither a direct, verifiable experience nor a known personal fact like one's address or age. It is a global retrospective judgment, which in most cases is constructed only when asked and is determined in part by the respondent's current mood and memory, and by the immediate context. In an elegant demonstration of the power of context, Schwarz (1987) invited subjects to the lab to fill out a questionnaire on life satisfaction. Before they answered the questionnaire, however, he asked them to photocopy a sheet of paper for him. A dime was placed on the copy machine for a randomly chosen half of the sample. Reported satisfaction with life was raised substantially by the discovery of the coin on the copy machine-clearly not an income effect. Other research indicates that reported life satisfaction is influenced by the current weather (higher on nicer days); although if individuals are first asked explicitly about the weather, the weather does not influence their reported life satisfaction In addition to being affected by researcher initiated manipulations of context Considerations of the effects of context, mood and duration neglect indicate certain limits on the reliability of the standard life satisfaction and happiness questions, but they are not necessarily grounds for dismissing the method altogether. The idiosyncratic effects of recent, irrelevant events are likely to average out in representative population samples. Moreover, research finds that retrospective evaluations are relevant for some subsequent choices, so measures of satisfaction may be relevant for future decisions despite their shortcomings as a measure of real-time experience. For example, job satisfaction is a strong predictor of workers' subsequent turnover

We use randomized grants to generate shocks to capital stock for a set of Sri Lankan microenterprises. We find the average real return to capital in these enterprises is 4.6%–5.3 % per year), substantially higher than market interest rates. We then examine the heterogeneity of treatment effects. Returns are found to vary with entrepreneurial ability and with household wealth, but not to vary with measures of risk aversion or uncertainty. Treatment impacts are also significantly larger for enterprises owned by males; indeed, we find no positive return in enterprises owned by females. I.

The Moving to Opportunity (MTO) demonstration assigned housing vouchers via random lottery to public housing residents in five cities. We use the exogenous variation in residential locations generated by MTO to estimate neighborhood effects on youth crime and delinquency. The offer to relocate to lower-poverty areas reduces arrests among female youth for violent and property crimes, rela-tive to a control group. For males the offer to relocate reduces arrests for violent crime, at least in the short run, but increases problem behaviors and property crime arrests. The gender difference in treatment effects seems to reflect differ-ences in how male and female youths from disadvantaged backgrounds adapt and respond to similar new neighborhood environments. I.

Abstract not found

The view that the returns to public educational investments are highest for early childhood interventions stems primarily from several influential randomized trials- Abecedarian, Perry, and the Early Training Project- that point to super-normal returns to preschool interventions. This paper implements a unified statistical framework to present a de novo analysis of these experiments, focusing on two core issues that have received little attention in previous analyses: treatment effect heterogeneity by gender and over-rejection of the null hypothesis due to multiple inference. The primary finding of this reanalysis is that girls garnered substantial shortand long-term benefits from the interventions. However, there were no significant long-term benefits for boys. These conclusions would not be apparent when using ”naive ” estimators that do not adjust for multiple inference.

In Project STAR, 11,571 students in Tennessee and their teachers were randomly assigned to different classrooms within their schools from kindergarten to third grade. This paper evaluates the long-term impacts of STAR using administrative records. We obtain five results. First, kindergarten test scores are highly correlated with outcomes such as earnings at age 27, college attendance, home ownership, and retirement savings. Second, students in small classes are significantly more likely to attend college and exhibit improvements on other outcomes. Class size does not have a significant effect on earnings at age 27, but this effect is imprecisely estimated. Third, students who had a more experienced teacher in kindergarten have higher earnings. Fourth, an analysis of variance reveals significant classroom effects on earnings. Students who were randomly assigned to higher quality classrooms in grades K-3 —as measured by classmates’endof-class test scores —have higher earnings, college attendance rates, and other outcomes. Finally, the effects of class quality fade out on test scores in later grades but gains in non-cognitive measures persist.

Abstract not found

Abstract not found

Uncovering the effect of school racial composition is difficult because racial mixing is not accidental but instead an outcome of government and family choices. Using rich panel data on the achievement of Texas students, we disentangle racial composition effects from other aspects of school quality and from differences in abilities and family background. The estimates strongly indicate that a higher percentage of black schoolmates reduces achievement for blacks, while it implies a much smaller and generally insignificant effect on whites. These reJohn F. Kain fully participated in this research, but sadly he died before its publication. An early version of this article was presented at the Brookings Conference on Empirics of Social Interactions (January 2000). Our thanks to conference participants, David Armor, Phil Cook, Jonah Gelbach, Caroline Hoxby, and Jens Ludwig for helpful comments. Support for this work has been provided by the Spencer Foundation, the Mellon Foundation, the Smith Richardson Foundation, and the Packard Humanities Institute. Contact the corresponding author, Eric A. Hanushek, at hanushek@stanford.edu. 350 Hanushek et al. sults suggest that existing levels of segregation in Texas explain a small but meaningful portion of the racial achievement gap.

Abstract: Strengthening the relationship of accountability between health service providers and citizens is by many people viewed as critical for improving access to and quality of health care. How this is to be achieved, and whether it works, however, remain open questions. This paper presents a randomized …eld experiment on increasing community-based monitoring. As communities began to more extensively monitor the provider, both the quality and quantity of health service provision improved. One year into the program, we …nd large increases in utilization, signi…cant weight-for-age z-score gains of infants, and markedly lower deaths among children. The …ndings on sta ¤ behavior suggest that the improvements in quality and quantity of health service delivery resulted from an increased e¤ort by the sta ¤ to serve the community. Overall, the results suggest that community monitoring can play an important role in improving service delivery when traditional top-down supervision is ine¤ective. This project is a collaborative exercise involving many people. Foremost, we are deeply indebted to Frances Nsonzi and Ritva Reinikka for their contributions at all stages of the project. We would also like to acknowledge the important contributions of Gibwa Kajubi, Abel Ojoo, Anthony Wasswa, James